Skipper's dashboard

This is an interesting idea. – InfernalMachine, on September 06, 2008 15:24

I don't think we can base question quality on users or H$ alone. For example, some questions have awful suspend/settlement times which makes it easy to gain a little dough if you time it right. For example: "Will company X's stock reach $50/share by the end of the month? Suspends: last day @ 11:59pm". We've all done it - it's after (the real live) market closing so you dump loads of change into the question. The return isn't good but it's safe and pays out in a couple of hours. Those questions have tons of users and tons of H$ but they're not really quality questions.

I was thinking about this some and came up with an off-the -wall solution that defeats this problem. Ready? Base the quality on users or H$ bet on the wrong answer(s). Crazy. Hear me out. If the market settles at 99% that's a bad question with a bunch of people obviously gaming the system. In that case the 1% voting on the wrong answer would tip you off and the quality reward should stink appropriately. However if the market closes at 60% that's a pretty good question and the 40% multiplied by users and/or H$ should reward accordingly.

This would reward people for creating popular, quality questions.

I like what Markov said: "co-mingling this [day-trading] with the vast majority of other question scenarios on Hubdub seems to be like mixing oil and water" and "...have a separate Hubdub for it that does not registers earning towards the primary Hubdub site."

I think "separation" is the only way to please both factions: the ordinary, lightly involved majority, and the heavily involved in-game day-traders. However, there should not be a separate Hubdub site, but only a separate assets "counter" for day-trading winnings, and a separate leaderboard for winnings on one-day stock market predictions.

This separate leaderboard could also accommodate in-game betting on sporting events, which I suspect would be very popular.

Hey everyone, Ive been reading this forum and I just wanted to say that rewarding the question creators is something I always thought should be a part of the site.
I think one way to do it would be to reward the creator at by the number of correctly settled questions. Maybe every time you reach a certain "level" throw in a bonus.
An example:
beginner
level 1 20 questions settled (and $H500 bonus)
level 2 50 questions ($H1,000 bonus)
level 3 100 questions ($H2,500 bonus)
level 4 200 questions ($H5,000 bonus)
level 5 500 questions ($H10,000 bonus)
You can have an icon similar to Super user indicating what question creation level each user is.
If you stress the fact that the questions have to be settled to count, this will help encourage quality questions.

Good suggestion Markov - thanks. We'll run some experiments to see how these look. One thing to bear in mind is that we have to balance combating uninteresting questions with encouraging niche questions that may not be well represented on Hubdub (and also may attract new members).

Interesting idea. – InfernalMachine, on August 19, 2008 00:48

@Tom - How about "Average # of Individual Predictors per Question"? This would encourage interesting questions while simultaneously discouraging people from creating hundreds of 'narrow interest' questions just to get their total number up. Personally I would rather see people filling the question slots with 5-10 interesting question submissions per month rather than 200 questions about how fast some celebrity's toenails are growing. The latter simply serves to knock the interesting questions out of prominence and thereby reduces the number of people who see and wager on them. Maybe a minimum of 5 questions would have to be asked to qualify for question ranking and the ranking could be reset on the first of each month. Downplaying quantity to encourage quality would be a step in the right direction.

@Markov - good points - I think you're right, the reward for question creators should be based more on interest than number created (although it's good to track that too). We were thinking basing it on the number of individual predictors per question, added across all questions.

It's on the list of things we intend to do, but it's a long list so it probably won't happen for a month or so

Good question, Bayoubear! I've been trying to figure this out myself. I noticed my site visits dropped off quite a bit a few weeks ago and realized I was losing interest in the questions in general. I certainly may stand corrected, but it seems over time a relatively small number of users were asking a disproportionate number of questions. That is all cool and kudos to them for doing it. However, it seemed to me to have precipitated two things: 1) a number of people were asking questions just for the sake of asking a huge number of questions, many of which had a rather narrow interest, and 2) the vast majority of users were not asking any questions at all for some reason. Why this may have been happening, I do not know. Somebody suggested adding discussion forums. That is a great idea and might fuel interest from some of our 'silent' members. Also, having a question leaderboard based solely on the number of individual users who responded to a question, rather than H$ wagered. I sense the general populace is losing interest. Again, great question!

The first 12 hours criteria is good for nothing if there isn't any prediction made during that period. I think it should be strengthen by number of bids as well. – pembeci, on August 17, 2008 02:42

congrats! – Skipper, on August 16, 2008 22:51

At least the typo didn't result i 'pubic'..... – Bayoubear, on August 13, 2008 11:01

That's a nice idea. Obviously question creators are restricted for the first 12 hours.

In general question creators don't get called on to make a call on a question unless there is a dispute.

However, I like this. We are really packed on feature development but I will add this to my good ideas list.

I do see Bayoubear's point about new users. Did you realize that of the 9000+ names on the net worth leader board, 1000 have a $0 balance? Now some of this may be gaming-related, but it's possibly also players who give up when the money runs out, or at least just sign up for a new account. It would be a good idea to have a little incentive at the very start for a good question (one that settles), in the form of H$.

One thing we might do for question excellence is to invent a parallel currency like "lightbulbs". Many games have several types of thing you can or must collect (tiberium and gas from c&c for example). This way we could measure question quality & contribution without affecting the H$ as predicting-ability measure.

Makes sense to me! I'm also following the rule of thumb, which I posted in the other discussion: if settlement depends on any kind of interpretation or could be questionable, then I'm staying away from it.

Yes. Does seem like a reasonable idea.

I particularly like #2 & #3 of Tom's ideas.

I like this idea

I like all of the ideas given by Tom...my only concern about the money issue was to allow some earnings (particularly for newbies) to help players build up some funds....maybe limited the reward to the first five successful questions created...but if not, that's fine...just as long as we get more people involved in the creation of questions...

Thanks for the reply